How to bulk import redirects into Framer from a csv

In the project, open site settings → hosting → redirects. Each rule is an old path and a new path, and Framer’s redirects help article lists three things worth knowing before anything is imported:

• Only paths of the current domain. A rule redirects /old on the published domain. A whole old domain is redirected at its own host or dns provider, not here.
• Order is priority. Rules can be reordered by drag and drop, and the first one that matches wins. Specific rows go above the wildcards that would also catch them.
• Nothing follows a renamed page. If a page path or a CMS slug is changed later on the canvas or in the CMS, the redirect settings are not updated with it. Settle the new site’s urls before you build the map.

Changes take effect when you publish or update the site. Framer’s pricing page lists “site redirects” for the Pro plan and Enterprise, not for Free or Basic (as listed 3 october 2026); it states no maximum number of rules.

Framer’s bulk import article describes the columns, and the plugin’s open source parser shows exactly how a file is read:

The header is optional: the parser treats the first row as a header only when it reads exactly from,to, and otherwise imports it as data. A file with fewer than two or more than three columns is refused. A row whose from matches an existing redirect updates it; the import adds and updates but does not delete, so rules already on the site and not in the file stay where they are.

from,to
/about-us,/about
/leistungen/webdesign,/services/web-design
/news/spring-launch,/blog/spring-launch

That is the file Silentfrog writes when you pick plain path pairs (wix, hubspot, other) in the export: header from,to, one path pair per line, nothing else. Identity rows, unmatched rows and needs-review rows you did not verify are left out, so the file only holds redirects somebody can stand behind (the two downloads). There is no Framer entry in the picker, because the plain pairs already are the Framer file.

• Export what is there first. Open Redirect Sync from the plugins menu and choose export. The csv it writes is the backup, and it shows which old rules from an earlier redesign still exist.
• Import the map. Choose import in the same plugin and select the file. Each row becomes a redirect, or updates the one with the same from.
• Add the wildcard rules by hand in the redirects panel (below), and drag them under the individual rows.
• Publish. Nothing redirects until the site is published or updated.
• Test a few rows on the live domain with curl -sI: one redirect, then a 200.

Framer’s notation, from the redirects article:

* — matches any path, across slashes. /blog/* catches /blog/hello-world and /blog/2022/07/18/hello-world.

:1, :2 — in the target, what the first and second asterisk matched. /blog/*/post/* to /article/:1/entry/:2.

:slug, :year — a named segment between two slashes. /blog/:year/:month/:day matches /blog/2022/07/18.

Silentfrog writes Webflow’s (.*) and %1 for a directory rule ticked export as wildcard rule, and the plain pairs carry no wildcards at all: a directory rule is written out row by row there, so no url the crawl found is lost. That is safe to import as it is. If you want the shorter list, translate the rule into Framer’s spelling and add it by hand:

Then delete the rows the rule covers from the csv before importing, or keep them: a specific row above the wildcard gives the same answer, it is only longer. A wildcard also catches the old urls the crawl never found, which is the real reason to write one. Keep any row whose new slug differs from the old one above the wildcard, or the wildcard sends it to a slug that does not exist.

• Relative paths only. The from column must start with /. A csv of absolute urls from another tool needs the scheme and host cut off first.
• The final target, not a hop. If an old rule already sends /a to /b and the new map sends /b to /c, change the old rule to /c. Framer’s own Webflow migration guide says the same: send each old url to the closest relevant page and avoid chains.
• Locales. With localization on, expandToAllLocales decides whether a rule also applies under each locale’s prefix. It defaults to true; set it to false in a third column for rows that should not.
• Pages that kept their path need nothing. A redirect from a path to itself loops. Silentfrog never writes one; a hand-made list might.
• After publishing, request every old url once and expect one redirect, then a 200 on the new path. A 404 at the end is a target that does not exist; two redirects in a row is a chain.

tail -n +2 redirects.csv | cut -d, -f1 | while read p; do
  printf '%s ' "$p"; curl -s -o /dev/null -w '%{http_code} %{redirect_url}\n' "https://www.example.com$p"
done

Redirect Sync is strict about the shape and relaxed about everything else, so most redirect lists need one small change before they import:

The plugin reports a file it cannot read rather than importing half of it: “CSV had invalid fields” means the first row has fewer than two or more than three columns, “CSV had invalid row” means a later row is missing its from or to. Both are messages from the plugin’s parser, so they name the file problem, not a Framer limit.

Silentfrog’s generic csv is the one file not to import: it carries absolute urls, the match metadata and the unmatched rows, because it is the audit trail of the project. The redirects file next to it is the import.

• Framer redirects live under site settings → hosting → redirects; Pro plan or Enterprise.
• Redirect Sync, made by Framer, imports a from,to csv; from must be a relative path starting with /.
• An existing from is updated, nothing is deleted; export first as a backup.
• Wildcards are * with :1 in the target; add them by hand and keep specific rows above them.
• Silentfrog's plain path pairs export is already the file Redirect Sync reads.